Alteryx Designer Cloud Core cheat sheet

Input Data supports files (CSV, XLSX, JSON, Parquet), databases via drivers, and cloud storage connectors; the connector determines available options.

CSV delimiter, quote character, and first-row-as-header settings must be set at input; a wrong delimiter collapses all columns into one field.

Output Data file format is chosen by the file extension in the output path (e.g., .yxdb, .csv, .xlsx); the tool does not infer from the incoming schema.

The Alteryx .yxdb format stores field types and metadata and reads faster than CSV; use it for intermediate staging files.

Database inputs return a SQL query string; switching to a custom query lets you filter server-side and reduce transferred rows.

Field type mismatches at input (e.g., numeric-looking strings with commas) are resolved with a Select tool after input, not by editing the source.

Wildcard input (*.csv in the file path) reads multiple files with matching schema into a single stream; schema differences cause field nulls, not errors.

Relative output paths resolve against the workflow location; absolute paths break when the workflow is moved or run on a different worker.

Data Cleansing tool handles null replacement, whitespace trimming, case modification, and removal of null rows/columns in one pass.

Trim removes leading and trailing spaces only; internal double spaces require a Multi-Field Formula with REGEX_Replace.

Select tool changes field type, renames fields, reorders fields, and deselects fields; it is the standard fix for type errors.

Imputation tool fills nulls using a chosen method (mean, median, mode, or a fixed value) computed over the incoming field.

Formula tool creates or updates one field per expression; use Multi-Field Formula to apply one expression across many fields.

REGEX_Replace(Field, pattern, replacement) uses the Alteryx regex engine; backreferences in the replacement are written as $1, $2.

DateTimeParse converts a string to a date with an explicit format string; DateTimeFormat converts a date to a string with a format string.

Record ID tool adds a sequential integer field; it is commonly used to preserve original row order after sorts and joins.

Join tool outputs three streams: Left (unmatched left), Join (matched), and Right (unmatched right); all three anchors exist on every Join.

Join requires at least one key field on each side; keys must have matching types or the join silently produces no matches.

Join Multiple supports more than two inputs and requires selecting join type per input (inner or outer style).

Union tool stacks records; it matches fields by name and position depending on configuration and does not deduplicate rows.

Append Fields adds all fields from a source to every record of a target without a key; it multiplies rows if the source has more than one record.

Cross Join via Append Fields or Join with a constant key produces the Cartesian product of both inputs.

Summarize tool groups by selected fields and applies actions (Sum, Count, Count Distinct, Min, Max, Concat) on other fields.

Transpose pivots wide-to-long (field names become a Name column, values a Value column); Cross Tab pivots long-to-wide using a group, header, and value field.

The Filter tool splits data on a boolean expression into True and False outputs; both anchors must be handled or records are lost.

The Formula tool's conditional logic uses IF c THEN x ELSEIF c2 THEN y ELSE z ENDIF; nested IFs require matching ENDIF keywords.

The Multi-Row Formula tool references prior or subsequent rows via Row-1, Row+1, or Row:0 for running totals and comparisons.

The Running Total and Running Average tools accumulate values in record order; sort before them if order matters.

The Sample tool can take the first N, last N, every N, or a random percentage of records; it is deterministic except for random modes.

The Sort tool orders records by one or more fields with ascending/descending per field; it does not change field types.

The Unique tool deduplicates by selected key fields and emits Unique and Duplicate outputs; the first record per key is kept as unique.

The Test tool validates a condition and routes failing records to an error output; use it to enforce data-quality gates.

The Reporting category builds presentation output (Tables, Charts, Layouts, Render); it does not modify the data stream for downstream tools.

The Table tool converts incoming records into a formatted table; field selection and grouping determine the table structure.

The Render tool produces a finished report file (e.g., PDF, PNG, XLSX) and requires a Layout or table/chart input.

The Chart tool creates visualizations (bar, line, pie, scatter) from summary data; numeric axes require numeric field types.

The Layout tool arranges report elements vertically or horizontally; without it, multiple report pieces may render separately.

The Summarize tool is the standard pre-aggregation step before charts and tables; charts built on unaggregated rows misrepresent values.

The Browse tool (or Results pane) inspects data at any anchor during development; it is a debugging aid, not a data-transformation step.

Report output to Excel can place a table on a named sheet; PDF output requires the Render tool, not Output Data alone.

Designer Cloud separates authoring (browser-based canvas) from execution on cloud compute; workflows run server-side, not on the local machine.

Connections to data sources are stored as governed connections in the platform; users reference the connection rather than embedding credentials.

Role-based access controls determine who can view, edit, run, or share workflows and who can administer connections and users.

Workflows and datasets are organized in a shared workspace; permissions inherit from workspace settings unless overridden per item.

Data lineage records which inputs, transformations, and outputs a workflow touches, supporting audit and impact analysis.

Scheduling runs workflows on a defined cadence using platform credentials; a scheduled run fails if the referenced connection is unavailable to the owner.

Version history retains prior workflow states; restoring a version reverts the canvas without altering source data.

Publishing or sharing a workflow exposes it to other users; embedded credentials and personal connections are the usual cause of run failures for other users.

A red anchor or tool error indicates an invalid configuration or schema mismatch; hover the tool to read the specific message.

'Field not found' errors usually follow a rename or a Select-tool change upstream; trace field names back through the stream.

Join producing zero matches commonly means mismatched key types or hidden whitespace; trim and align types before joining.

Performance: push filters to the database query, avoid reading unneeded fields, and prefer Select early to shrink the data stream.

The Cache and Run Workflow option reuses upstream results during iterative development; disable it before final validation.

Block Until Done controls execution order when multiple output branches must run sequentially (e.g., write a file before reading it).

The Performance Profiler records per-tool execution time; use it to locate the slowest tools in a long workflow.

A workflow that runs locally but fails in the cloud usually depends on a local file path, a local driver, or an unshared connection.